Characterization of a novel superconducting imaging surface magnetoencephalography system.

摘要

A novel system is under development at Los Alamos National Laboratory (LANL) to measure weak magnetic fields created by neuronal electrical activity in the brain. Commercial magnetoencephalography (MEG) systems exist to measure these fields. However, the LANL system incorporates a novel superconducting imaging surface (SIS) around the sensors. This SIS blocks background noise from the sensors, thus reducing the need for an expensive shielded room while improving performance.; The presence of the superconductor surface introduces an additional magnetic field in addition to the one that is created by the neuronal source. The distortion of the primary field must be fully understood and characterized in order for the LANL system to accurately localize any sources. A magnetic field integral equation is derived that gives the field due to a current density source in the presence of a superconducting surface. For more efficient calculation, a numerical model based on the finite element method (FEM) was also developed. This model was tested for accuracy against theory, in the case of simple geometries, and experiments, in the case of the complex shaped helmet superconductor.; The performance of the LANL SIS MEG system will eventually be benchmarked with a commercial MEG system, specifically the Neuromag-122. The gradiometer like performance of the LANL magnetometers in the presence of the SIS is analyzed. Experiments using artifical and human sources were conducted. These sources were localized using algorithms with similar protocols. Consequently, no commercial software was used. In the Neuromag-122 case, MRI and head tracking data were also collected and utilized in the localizations.; Single trial tools were also used to analyze the human data in the LANL SIS MEG case. These tools included the use of a blind source separation algorithm called Second Order Blind Identification (SOBI) and a novel use of wavelets for denoising in the case of low signal to noise.